Mojifall: an alternative visualization of Emojitracker's data

[axelpale]

First of all, thanks for your work on Emojitracker. I want to let you know about a project of mine that uses the data collected by Emojitracker and that could be fun for you to know about, at least.

So, what I have worked on is an alternative visualization of the emoji score chart. As comparing large numbers is a bit hard for people, and because bar graphs are boring, this viz aims to represent the frequency distribution of emojis as a waterfall, or mojifall if you may. There is already a demo available at:

demo.akselipalen.com/mojifall-openmoji-emojitracker/

For a bit of background, the project started as an approach to display the variety of great work done by OpenMoji organization. Soon the project found an interest to reflect the real-world use of emojis and we chose Emojitracker to be a suitable data source. For details about the development this far, see OpenMoji Issue #30.

Currently Mojifall reads a dump of emojitracker's data I downloaded some days ago via emojitracker-rest-api. In other words, it does not consume the API directly. However, I am interested to develop Mojifall further to visualize the current use of emojis, for example over the last 24 hours (see related issue #37). My motive here, in addition to being an open-source enthusiast and OpenMoji patron, is to find good real-world use cases for Sprinkler.js, an open-source stream-visualization lib I have been baking. Usage of emojis might be one of those.

Any thoughts? Also, if you feel Emojitracker could benefit from Mojifall in a way or another, I am happy to help.

[Pr0methean]

This is cool. It'd be even better if the frequency of emoji could be normalized to the long-term average rate, so that those used recently more than usual would appear the most frequently.

[axelpale]

@Pr0methean Thanks. You mean that by normalizing a short-term rate distribution to the long-term average, we could inspect how the current use of emojis differs from the general. That would be interesting.

If only there be an API for querying the recent usage, say during last 24 h or 1 week. On the other hand I could hook on the real-time emoji stream to get the recent trend. However, writing that requires seemingly large amount of work when compared to just having two static distributions, one short-term and one long-term, to play with.